Mars (Liyao) Gao

Ph.D. student
Paul G. Allen School of Computer Science & Engineering
University of Washington

GitHub Google Scholar

About me

I am a Ph.D. student in the Paul G. Allen School of Computer Science & Engineering at the University of Washington, advised by Professor J. Nathan Kutz. My research lies in AI for scientific discovery, with a focus on developing interpretable and generalizable learning frameworks for complex spatiotemporal and dynamical systems. I work at the intersection of symbolic regression, sparse modeling, and deep learning, aiming to uncover governing equations and enable reliable long-term prediction to accelerate scientific understanding. My long-term goal is to build robust machine learning methods that can bridge data and physical laws across domains like physics, climate science, fluid dynamics, neuroscience, and materials science. I am broadly interested in deep learning, statistical learning theory, Bayesian methods, time-series modeling, scientific computing, and recently agentic flow for science.

News

Selected publications

Sparse identification of nonlinear dynamics and Koopman operators with Shallow Recurrent Decoder Networks. Mars L. Gao, Jan P. Williams, J. Nathan Kutz. [GitHub] [Colab] [Website] [Youtube] arXiv, 2025.

Bayesian autoencoders for data-driven discovery of coordinates, governing equations and fundamental constants. Mars L. Gao, J. Nathan Kutz. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences.

Convergence of uncertainty estimates in ensemble and Bayesian sparse model discovery. Mars L. Gao, Urban Fasel, Steven L. Brunton, J. Nathan Kutz. In submission.

Deformation Robust Roto-Scale-Translation Equivariant CNNs. Mars L. Gao, Wei Zhu, Guang Lin. Transaction of Machine Learning Research.

On Optimal Early Stopping: Over-informative versus Under-informative Parametrization. Ruoqi Shen, Mars L. Gao, Yi-An Ma. arxiv.

Contact

Email: marsgao [at] uw [dot] edu